Turning investments green in bond markets: Qualification, devices and morality

Abstract

This paper explores the issuance and growth of transition and sustainability-linked bonds into the green market segment normally reserved for green bonds between 2018 and 2021. Using a performative economics and STS approach we examine how key terms within environmental political discourse – transition, green, sustainability – have been incorporated into the operation of investment markets, and given specific, if unstable and contested, technico-economic forms. Using event ethnography, industry literature and primary interviews, we examine how classification works as a market device, by exploring why the newer ‘transition bond’ was favoured by some investors but not others in comparison to green and sustainability bonds. We argue that the credibility of different green labelled products is being mediated by uneven references to scientific evidence, in the context of competing marketization strategies, which have world-making effects.

Keywords:

• green bonds
• transition bonds
• sustainability-linked bonds
• green finance
• marketization
• science and technology studies (STS)

Introduction

Since the 1990s, mobilizing and redirecting capital to facilitate transitions toward a lower carbon economy has grown as a central tenet of climate change governance. The centrality of the private sector was underscored by the Paris Agreement (2015) which committed to ‘Making finance flows consistent with a pathway towards low greenhouse gas emissions and climate-resilient development’ (Article 2). But green markets are young in historical terms, a feature of the post-2008 financial crisis world, and arguably ill-equipped to meet this challenge at scale. The first green bonds date from the 2007 and 2008 issuances by the World Bank and the European Investment Bank (EIB) respectively and in widely used ‘Green Bonds Principles’ are defined as ‘any type of bond instrument where the proceeds will be exclusively applied to finance or re-finance, in part or in full, new and/or existing eligible green projects’ (International Capital Markets Association [ICMA], 2021). From the mid-2010s the market started to generate a group of products which have both differences and similarities in name and purpose, such as transition bonds, social, sustainability and sustainability-linked bonds (SLBs). For example, transition bonds, or climate transition bonds, can be ‘used by companies in carbon-intensive sectors to help fund credible net-zero transition strategies’ (London Stock Exchange, 2022). Meanwhile, SLBs are increasingly common in European markets, and present as something of an alternative to green and transition bonds. SLBs have a flexible approach to signalling green using internal and processual key performance indicators (KPIs), rather than the green and transition bonds’ earmarking of a ‘use of proceeds’ activity directly.

In 2020 and 2021, repeated quantitative easing in major markets, combined with extremely low interest rates and reduced yields in vanilla (normal) bonds, prompted growth in the volume and size of the green finance market. Although still negligible compared to the total universe of fixed-income products, total green, social, sustainability and sustainability-linked (GSSS) bond issuance reached $1.03 trillion in 2021, over 69 per cent higher than the $606 billion of 2020, and more than triple the $326 billion of 2019 (EF, 2022a, p. 1). A ‘sovereign super-charger’ effect (EF-curated webinar, July 2021, S8) also contributed to growth, with the EU alone recording five of the top 10 annual issues by size in 2021 (EF, 2022a, p. 2; see Hinsche, 2021). There also appeared to be considerable excess investor demand for green bonds. For example, the September and October 2021 UK Treasury Green Gilt (Bond) issues (collectively worth £15 billion) were over 12 times oversubscribed (S15, March 2022). Most market commentators in 2022 believed that green bonds were expanding, transition bonds were faltering, and SLBs were the new rising star. But we have found outliers that suggest that a different product landscape could still emerge. For example, the London Stock Exchange, spurred on by the success of its Green Economy Mark and Sustainable Bond market segment, launched a specific segment for climate transition bonds in 2021 (London Stock Exchange, 2022), suggesting a growing optimism for transition bonds, despite many commentators predicting their decline.

In this context, this paper is concerned with the unstable and contested distinction between the classificatory technologies of ‘green’, ‘transition’ and ‘sustainability’ in investment bonds. It describes and analyses the ways in which the classification process works as a ‘market device’ to constrain or expand issuance in the three categories, providing contested meanings for each label. It explores recent product fragmentation, asking whether this signifies market failure and fracturing, or the differentiation, innovation and imagination required for a successful green finance market. Afterall, green finance warrants attention, not just because of its growing overall size, but because of the hopes people have that green finance can avert an existential planetary climate crisis. Understanding market-making will assist us in aligning finance to a healthier climate future. Our approach is inspired by performative economists and science and technology studies (STS) which predict that market making will be characterized by iterative experiments where inductive processes generate products that ‘work’, others that do not, and yet more that emerge from partial outcomes, where diversification provides bespoke products for specific investors in a process of continuous looped experiments.

This paper describes how the green market is still forming, showing that these experiments in product design are part of that process, and that final classification outcomes are unclear. However, we also detail underlying and foundational conundrums that all green products currently embody, which have wider theoretical significance. We note that scientific verification of the claims of green products is both expensive and intellectually complex for financial institutions; is not regularly used in the immanent valuation process for GSSS products (see Langley et al., 2021, on the difference between mainstream and ethical qualification of assets); but that science is still regularly conjured in additionality narratives that perform verification for green labels. We argue, following others, that the negotiation between cost, the veracity of scientific knowledge claims, and financial returns remains a space of tension for issuers and investors, but that the issue of just how much is enough science to give authority to green credentials, while simultaneously ensuring profitability, is emerging as the key node of contest, and is worthy of further research. The central contribution of this paper is to show that the classification processes which pacify green products, also reflect and contribute to wider, fundamental disagreements over how transition to lower carbon emissions is to be societally defined, managed and financially resourced. When issuers and investors consider these questions further, transition bonds might ‘rise’ again, as they have a design affinity with increasingly popular forms of scientific future-making in transition, scenario and pathway modelling. On the other hand, SLBs are more easily scalable, side-stepping the hoary relationship between assets and their carbon properties, effectively deferring consideration of the planetary limits to growth. In short, the significant efforts in metrology described here are yet to yield immanent changes within asset valuation processes: markets remain in stasis with only cheap and contested signifiers of green, not helped by the demise of the transition product which promised change pegged to temporally progressive investment decisions. However, there is qualitative evidence here that these narrative contests are changing the allocative decisions of investors, ceteris paribus, which will affect bond costs and resource allocation in climate change governance in the long-run.

Green and transition bonds in market experiments

Our examination of the classification of green, transition and SLBs owes an intellectual debt to three main bodies of work. First, studies in performative economics and the sociology of markets which have drawn attention to how markets and commodities are created in transitive socio-technical processes; (e.g. Asdal & Cointe, 2021; Callon & Muniesa, 2005; Çalıskan & Callon, 2009, 2010); actively and relationally constructed, ‘enacted rather than observed’ within ‘a diversity of practices’ (Mol, 1999, p. 77). To understand how specific products are ‘made’, discursively and materially, we evaluate the ‘collective process’ of qualification (Bracking et al., 2018; Callon et al., 2002, p. 203), as various actors do ‘metrological work’ with ‘heavy investments in measuring equipment’ (Beckert & Musselin, 2013; Çalıskan & Callon, 2009; Callon et al., 2002, p. 199; Langley et al., 2021, pp. 497–500). The work of Muniesa et al. (2007) on ‘market devices’ (p. 2) is used to analyse the various classification and category-making processes pursued by issuers and investors, as transition, green and sustainability bond products are qualified, certified and guaranteed (see Çalışkan & Callon, 2010, pp. 7–8).

Second, we draw on STS studies, an ontologically closely related body of work, which has explored the construction of knowledge-claims (e.g. Latour, 2004) and the devices created by scientists, in different disciplines, to organize knowledge and objects in their respective fields, such as Linnaeus's classification of species (see Müller-Wille & Charmantier, 2012), as well as the concept of ‘species’ itself. A broad range of related taxonomies exist to classify different bits of the environment, ranging from rocks and minerals to ecosystems, and, controversially, people (see Bowker & Star, 2000). Two insights from this work are particularly relevant for us: first, that these taxonomies are always changing, being culturally and historically situated – and therefore ‘constructed’ and not given; and second, that they contribute to different forms of socio-political arrangements (e.g. Gieryn, 1983; Haraway, 1988; Shapin, 1998). This entails that classification and taxonomies are not neutral to the organization of knowledge, or of societies: for example, the concept of species serves to articulate regions which are the most biodiverse, and guide conservation decisions over which areas should be prioritized. In Sorting things out, Bowker and Star (2000) show how classifications, i.e. ‘a spatial, temporal, or spatio-temporal segmentation of the world’ (p. 10) are pervasive, yet often taken for granted and ‘invisible’. Yet, classification systems render voices visible or invisible, often reflecting and reinforcing existing power relations and hierarchies (e.g. Esguerra, 2017; Konefal & Hatanaka, 2011). In a similar way, deciding on what counts as green [boundary-making], and the relations between categories of different types of green product, require similar metrological and taxonomic efforts.

Third, recent work within economic anthropology, inspired by Foucault (2005), has shown how speculative investments are justified within narratives of green worth and value (Bear, 2015, 2020; Leins, 2020, 2022). Bear (2020) summarized that speculation is ‘future-oriented affective, physical and intellectual labour that aims to accumulate capital for various ends’ (p. 2), a ‘technology of imagination’ (Bear, 2015, 2020). Meanwhile, Leins (2022) studied financial analysts and showed how they used feelings, individualized strategies, imagination and simple storytelling to make explanatory narratives about future risk – converting speculation into ‘investment’. Similarly, in a case study of Gothenburg (Sweden), the first city to issue municipal green bonds, García-Lamarca and Ullström (2022) draw attention to the role of affects and emotions in facilitating the proliferation of green bonds as parts of a ‘feel good economy’, reflective of a post-political discourse which rejects antagonism and conflict and promotes ‘triple-win’ situations. These authors, and others (e.g. Ghosh, 2020; Mellor & Shilling, 2017), show how green products are not just about economic logic but also about ‘morality’ and people wanting to feel positive emotion (c.f. the ‘green halo’ effect for food labelling, Sörqvist et al., 2015). This moral story making in green bond markets is contributed by a range of agencies, verifiers, institutions and climate scientists in the process of green bond issuance (Tripathy, 2017), in an ‘ethical liberal modality of qualification’ which appeals to the social values of investors (Langley et al., 2021, p. 505). We have also been informed by scholarship within psychoanalytical studies and subjectivity (cf. Zizek) exploring why actors continue to develop products, even if they are aware that they do not perform well or do not work (e.g. on the ‘fantasy of carbon offset’ see Watt, 2021). Under these perspectives, humans are treated not as rational homo economicus but as affective beings (see also García-Lamarca & Ullström, 2022).

This paper also draws from an expanding literature, from various traditions, on how ‘green’ is being defined in relation to standards (for example, Berrou et al., 2019; Nedopil et al., 2021; Perkins, 2021; Torsten & Packer, 2017), brands and branding as a technology of government (Bryant, 2013), disclosure (Ameli et al., 2020; Griffin & Jaffe, 2022; Troeger & Steuer, 2021), taxonomies, rating and reporting (e.g. Cox, 2022), more or less specific carbon properties (Langley et al., 2021) and regulatory initiatives and governance (Bracking & Leffel, 2021). In terms of classification, Perkins (2021) argued that the emergence and diffusion of green bonds was conditioned by a contested certification process, with the [more flexible] green bond principles achieving dominance over the climate bond standards. Meanwhile, in findings that resonate with our own (below) Langley et al. (2021) noted that the classification of ‘low-’ or ‘high-carbon’ properties in assets was ‘contingent, contested and compromised, featuring contrasting modalities of qualification that are decarbonizing capital in uncertain and incomplete ways’ (p. 494).

Drawing on these etymologies and studies, this paper focuses on how specific product framings – green, transition, sustainability – affect the overall green bond market. Looking at transition bonds, a relatively recent and understudied object, adds to the extant literature on green finance. We focus on meaning-making: how are products constructed, and how are they classified? How is the socio-technical order of green, transition and sustainability bonds emerging and mutating, fracturing or stabilizing? Why did some actors advocate for transition bonds, while others resisted and maintained that more generic categories were sufficient?

Methodology

Our results are informed by qualitative research, framed by three empirical research questions, conducted by the authors between October 2020 and October 2021. First, what are transition bonds and how do they work? Second, how are transition bonds different from other bonds? Third, why are issuers producing new classifications and with what consequences for green finance markets? We combined a literature review with primary data collection, focusing on processes of qualification, calculation and morality in product and market-making. The review included financial journalism, grey (corporate) and academic literature using the primary search terms ‘transition bond’ and ‘transition bond + framework OR guideline’ and ‘transition bond + [issuing company e.g. Mafrig]’, repeated with ‘green’ and ‘sustainability linked bonds’. We then conducted expert elicitation and expert semi-structured interviews with transition bonds issuers, investors, ‘brown’¹ target companies and regulatory institutions, and multi-sited event ethnography. Because of the COVID-19 pandemic, our event ethnography was conducted remotely, with the authors attending webinars and conferences which were recorded and later transcribed. This exploited a higher, COVID-related, temporary openness within these epistemic communities. In total we interviewed 19 senior market participants, financial investors, regulators and issuers (coded I 1–19), and attended 14 specialist online events producing thematically coded transcriptions (summarized and coded S1–14).

While there is some literature on green and climate finance, our study is unique in its comparative analysis of transition bonds, and because of our conceptual analysis of qualification, calculative devices and moral narrative, rooted in the politics of classification. The next section reviews existing academic and grey literature, while the fifth section presents the ethnographic and interview material, informed by these results.

The emergence of transition bonds

The origins of transition bonds are found in instruments used by US electric utility companies in the 1990s, who were given permission to levy customers after regulatory tightening to recover stranded or sunk costs. More recently, the high-emitting firms Castle Peak (Hong Kong power producer, 2017), SNAM (Italian energy operator, 2019) and Mafrig (Brazilian beef producer, 2019), used transition bonds for future-oriented projects, the latter for cattle purchases from the Amazon Biome differentiated by a higher standard of supplier ESG. But despite these early offerings, transition bonds did not expand as rapidly as expected, with only 11 issued in 2020, 12 in 2021, and 54 in 2022. The latter figure looks promising, but the overall value was only US$3.5 billion of a total green finance market of US$863.4 billion (CBI, 2023), and discussion about the failure of this instrument to go ‘mainstream’ (Ng, 2022; Webb, 2022) has prevailed, despite much effort in classification and regulatory taxonomies.

Transition bond market-making

The development of transition bonds has been non-linear, with a few large issues by actors in high emitting industries. Other actors, financial institutions, industry associations, investors, think tanks and governments, have also played a significant role in market expansion through the development of ‘transition’ guidelines and taxonomies. Most transition bond issues are accompanied by bespoke issuer level bond guidelines, but many third-party proposals have also attempted to detach, define and standardize ‘transition’ and ‘transition bond’, in ‘prescriptive approaches like taxonomies, to less prescriptive ones like guidance and principles’ (Tandon, 2021, p. 20). All converge around emission reduction using transition finance to fund assets, operating and capital expenditures (Tandon, 2021, p. 21).

One of the first to propose new transition bond guidance was AXA Insurance in 2019. They argued for a ‘new type of bond, distinct from green bonds’ for companies ‘which are “brown” today but have the ambition to transition to green in future … i.e. most businesses in the world today’ (AXA Investment Managers (AXA), 2019). Transition bonds were defined flexibly as:

Any type of bond instrument where the proceeds will be exclusively used to finance fully, or partly, or refinance new and/or existing eligible transition projects. (AXA, 2019)

They were for ‘industries which currently do not (and for the foreseeable future may not) have sufficient green assets to finance but do have financing needs to reduce their greenhouse gas footprint’ (AXA, 2019).

In an early recognition of potential misuse of the label, AXA insisted on an environmental strategy to ensure that the transition project was ‘material to the business’ (AXA, 2019). BNP Paribas (2019), also publishing guidelines in 2019, made a similar, but starker, distinction between green and transition bonds where the former should be for ‘green industries alone’. Similarly, guidelines from the Research Institute for Environmental Finance (REIF) (2020), a Japanese non-profit, argued that transition finance should be for high emitting firms, thus drawing a distinction between green and ‘brown’ based on the general characterization of the issuer. Thus, for AXA, BNP Paribas and REIF, transition bonds were defined as an alternative to green bonds, which solve the ‘pipeline limits’ (not enough ‘green assets’ to finance) of green bonds. They helped produce an enduring precedent, that finance would attach to a ring-fenced ‘transition project’, in an otherwise high-emitting industry or issuer. This idea is now salient to definitions of transition bonds, despite that there is a fuzzy tension or overlap with activities that green bonds could also cover, and in terms of how far the ‘green project’ can in fact be a ‘green process’ or indeed merely a ‘green plan of intent’, at which point there are overlaps with SLBs.

The Climate Bonds Initiative (CBI) and Credit Suisse published influential guidelines in September 2020, which defined a wider concept of ‘transition’, applicable at both ‘activity and entity level’, with five categories of economic activities that issuers could use to determine the appropriateness of a ‘transition label’: (1) near zero, (2) pathway to zero, (3) no pathway to zero, (4) interim and (5) stranded. The guidelines state that there are no ‘labels applicable at an entity level’ for the ‘no pathway to zero’ and ‘stranded’ economic activities (CBI, 2020, p. 5), thus excluding stranded assets since these cannot be brought ‘into line with global warming targets’, in a pathway to the post-2050 economy (CBI, 2020, pp. 4, 7). The CBI (2020) further argued that ‘credible transition goals’ must be in alignment with the Paris Agreement at 1.5°C as opposed to Nationally Defined Contributions (NDCs) which may not be ‘sufficiently ambitious’ (p. 13). This early work of identifying specific activities which would differentiate access to transition/green finance, proved salient to future taxonomy making, even though the CBI subsequently lost ground to the ‘more flexible’ Green Bond Principles (Perkins, 2021).

In contrast, the International Capital Markets Association (ICMA) adopted a wider more flexible transition concept in its guidelines of December 2020, which stated that ‘a corporate strategy’ addressing climate change risks should be ‘a pre-requisite to issuing a transition-labelled instrument’ (ICMA, 2020, p. 3), thus privileging the long-term planning of the issuing firm over the specific investable activity. Also, in contrast to the CBI, the ICMA accepted a level of geographical variability, stating that ‘transition pathways must be tailored to the sector and operating geographies of an issuer’ (ICMA, 2020, p. 2). As significantly they were reluctant to articulate a new ‘transition bond’ label, or definitions or taxonomies of transition projects, preferring to clarify ‘the issuer-level disclosures’ which would be necessary to use ‘transition’ within already existing green, social and sustainability-linked instruments (ICMA, 2020, p. 2). This was a surprise to market participants when the handbook was released (Furness, 2021), although the ICMA guidelines have become an industry standard for transition bonds nonetheless, because or despite of their flexibility.

At the national and supranational level there has also been fierce debate about how to define ‘transition’, taxonomies and appropriate data, with Japan, Malaysia, Russia, Singapore and Canada examples of first movers (see Tandon, 2021). The EU, via the Technical Expert Group of Sustainable Finance (TEG), incorporated transition bonds into the existing EU Taxonomy and Green Bond Standards (European Commission, 2020, p. 15) as activities which make a ‘substantial contribution to climate change mitigation’, but with caveats: they must have ‘(i) … greenhouse gas emission levels that correspond to the best performance in the sector or industry … (ii) … not hamper the development and deployment of low-carbon alternatives; and … (iii) … not lead to a lock-in of carbon-intensive assets’ (European Commission, 2020, p. 14), these latter providing for difficult evaluative and counterfactual considerations.

In sum, despite this proliferation of guidelines and taxonomies there remain basic definitional tensions: between expansive or limited definitions of transition; on whether the transition should be a unique activity or a process; over the types of indicators required to validate transition; as well as a boundary issue of deciding what is a stranded or irredeemably harmful asset which cannot transition. For example, the CBI considered Mafrig’s offer neither a legitimate green or transition bond as it did not directly address emissions, while Cadent gas company’s offer, notionally aligned with the EU taxonomy, was criticized for allowing carbon-intensive assets to be unnecessarily extended (CBI, 2020, pp. 7, 25). Meanwhile, a former Mirova portfolio manager said of the Mafrig $500 million issue, that ‘brown to light-brown is not enough’ (cited in Webb, 2022).

Despite these tensions, this definitional insistence on plausible pathways may yet be that aspect of transition bonds which gives them future dominance. This is because efforts to bring ‘more science’ into financial regulation – through paradigms such as climate risk disclosure reporting, Net Zero, Paris alignment, and ‘just transitions’ – assign new responsibilities which have commensurability with the future-making pathway aspect of the transition bond as compared with more static green bonds.

Standpoints of green finance investors and issuers

We thematically coded our interviews and event transcripts and four main contested modalities in the politics of classification emerged: (1) contested views on whether a new label was needed; (2) the disruptive emergence of SLBs; (3) metrology of transition and impact; and (4) investment decision-making. The next sub-sections cover these in turn.

Early market-making efforts to promote transition bonds were led by banking and investment executives, who believed that having two products could protect against the reputational damage of lending green finance to a high-emitting company, protecting the integrity of green bonds. For example, BNP Paribas (French International Bank) promoted this ‘brand-new’ label to ‘allow green bonds to retain their “purity”’ (BNP Paribas, 2019). Similarly, AXA investment managers advocated transition bonds for ‘brown’ companies with green transition ambitions, but with insufficient projects for ring-fencing in a ‘use of proceeds’ green bond (AXA, 2019). A certification expert explained the benefits of the transition label for reputational risk management:

Normally we’ve taken the stance that if the use of proceeds is green, it’s green, so it doesn’t really matter what the identity of the issuer is. But there has been one occasion where our certification board actually said no, they are such that they bring the idea into disrepute. (I17, July 2021)

Although he noted that this company may have been ‘unfairly lambasted’ since:

To be fair to them, I think that was actually part of their strategy to move away from brown … We want everyone to transition to green and obviously not everybody starts from being a pristine green. (I17)

However, beyond this common view on reputational risk, interview respondents did not agree on product categorization or on whether fixed-income environmental products should be diversified or standardized. We can identify five broad positions, the first of which best incorporates this original motivation to manage reputational risk: (1) transition bonds are substantially different to green bonds, a positive innovation that therefore requires different labelling; (2) transition and green bonds as similar products but still need distinctive branding to expand the investor pool and target specific investors; (3) transition bonds as just a subset of green bonds, so there is no need for a new label; (4) transition bonds are an unnecessary distraction that detrimentally restricts the growth of investor liquidity; and (5) the question of differentiation between green and transition bonds is becoming redundant as SLBs will displace both due to their relative efficiency benefits in issuance costs and monitoring.

Many experts concurred that the chief benefit of transition bonds would be expanded investor pools (position 2), to support high-emitting corporations with credible transition plans. For example, a chief sustainability officer at Credit Suisse commented that:

Today the green bond market is overweighted towards high-investment-grade sovereign, FIG and real estate issues, and there is not as much participation from corporates and lower rated credits. (S4)

However, there were detractors from this position that product diversification expands liquidity. For example, a Johannesburg Stock Exchange Board member noted that she:

would be surprised if anyone who is issuing a green bond wouldn’t be able to access the capital markets ordinarily … so it’s a marketing positioning exercise … which allows them to signal that this is the kind of company that we are. (I6, June 2021)

If true, green branding signals virtue, indirectly builds brand value and market share, may change the investor constituency, but may not contribute directly to expanding available liquidity.

Other actors argued that ‘green bonds’ were already very diverse, making ‘transition bonds’ unnecessary (position 3). For example, a representative of Nordea (Nordic Bank) wrote:

‘Transition’ is already baked into the green bond market. This is eloquently formulated in the Shades of Green methodology, where, in essence, anything that is not ‘dark green’ represents transitioning. Green bonds are/should be for everyone. To this point, it is somewhat hollow to say that an oil company can’t issue a green bond if you would buy a green bond from any of the major banks, most of whom have significant oil-related exposure on their balance sheet. (S1, November 2020)

Others agreed that even if the issuer is ‘brown’, if the use of proceeds case is credible on transition, then there is no need for further product differentiation. Our Stock Exchange Board member summarizes: ‘Why can’t they just issue a green bond? … green bonds have never been about the issuer – they’ve been about use of proceeds’ (I16). She accepts that people might be worried about greenwashing but responds:

So they want to know that their green bonds are pure and clean. But actually, the pure and green, clean, green bonds are not the stuff that really matters. What matters is getting the dirty stuff clean right? So we need to transition all this stuff. And so we need to finance it. (I16, June 2021)

She reminds us that investment bankers are paid structuring fees so differentiation may benefit them, but not issuers or users, arguing further that ‘inventing’ new categories for ‘the sake of it’ fragments liquidity (position 4), and is counterproductive to the expansion of green finance (I16). A BlackRock investor concurred that the issuance and monitoring costs for slightly different products would be prohibitive and just shrink the market (S14). In short, standardization expands liquidity.

Position 5, is that green and transition labelling is increasingly becoming redundant, because of several related arguments. First is that SLBs will take over from both (explored in the next section); second, that quality impact investors look beyond the label in any case; and third, that ‘green’ will become redundant as it is mainstreamed. For example, this impact investor is assessing impact independently of the bond label:

That’s not so important to our clients. What's important is that we are only investing in bonds that have positive impact. So, whether the bond has actually been given the green label or not does not determine whether we could invest in it. (Katie House, asset manager, Affirmative Investment Management)

Meanwhile, a senior analyst from a leading NGO certification organization noted that it was depressing that more assets were not green certified already:

We’d like to see that increase over time and there’s no reason why every asset shouldn’t have an adaptation resilience certification – I mean, why would you want to live in a building that isn’t [green certified]. (I17, July 2021)

He continues that electric cars and public trains as clearly examples of assets which have not requested a green label, but which have green credentials. Similarly, one ESG analyst imagined a future where the green bond disappears, leaving only ‘normal bonds’ for green companies, and ‘transition bonds’ for ‘brown companies’:

In time, as companies become greener, we can expect all corporate bonds issuance to be perceived as ‘green bonds’ and carry the same ‘greenium’ – without having to go through the administrative hassle or incur costs for registering a green bond. (I1, July 2021)

SLBs versus use-of-proceeds bonds

In a webinar curated by Environmental Finance in July 2021, staff analysts argued that the label ‘transition bonds’ had not really ‘taken off’. They continued that while this was partly attributable to the ICMA decision not to publish transition bond principles in their Climate transition finance handbook in December 2020, it was principally because of the benefits of flexibility (and potentially lower standards) attributed to SLBs. These latter were working well in ‘hard-to-abate’ sectors and for issuers previously excluded from the sustainable debt, fixed income universe, thus removing the ‘raison d'être’ of transition bonds and taking the ‘wind out of their sails’ (S8). The data supports this view. In 2021, while SLBs rose by more than nine times, transition bonds shrank to only nine issues of $4.4 billion (EF, 2022a, p. 4). Green bonds were similarly affected by competition from SLBs (and social bonds): they were a full 85 per cent of the total GSSS market in 2018, but just around half in 2020 and 2021 (EF, 2021, p. 3). But green bonds were predicted to remain more popular than transition bonds (EF, 2022a, p. 4).

Movements in market share are partly attributable to differences in product design between SLBs and fixed-income bonds. SLBs use key performance indicators (KPIs) rather than a use of proceeds narrative, where the KPIs are measured against sustainability performance targets (SPTs), which increasingly have performance linked step ups/step downs or ratcheting systems in interest rates paid on coupons. These ‘steps’ act as financial incentives to maintain sustainability outcomes: if an issuer does well against its targets, it pays less interest. Thus, SLBs are not simply an alternative to green and transition bonds, but potentially a substitute or competitor, where their popularity is attributed to their perceived advantages in terms of flexibility, implementation, management, cost and scalability relative to use of proceeds products. Data providers and strategists argue that SLBs have easier data management requirements than use of proceeds bonds, with the former using KPIs, SPTs and post-issuance reporting, but the latter sometimes involving multiple data points (sustainability data provider at the Luxembourg Stock Exchange, S6). This makes SLB KPIs: ‘much broader … much easier, and less granular … we don’t need to track where every dollar is going’ (I3).

However, limitations related to SLBs’ potential for ‘low ambition’ and measurability are also reported. One large asset manager warned on impact that, ‘We are a long-term investor, so we have a long-term institutional memory’, and issuers ‘need to keep [their] promises’ (S10), suggesting that sometimes they did not. In another webinar discussing SLBs in emerging markets, participants noted that while flexibility was an advantage, the ‘step up’ had to be ‘strategic’ and ‘material’, ‘not just a 1 per cent reduction of carbon in 20 years on a KPI’, and more than where ‘business as usual activities are reclassified and reallocated’ (S10). Many references to limitations were highly qualitative, which can partly be attributed to the design of SLBs: since SLBs rely on performance management frameworks, rather than verifiable ‘science’, flexibility and critique are both produced in qualitative narratives. But an emerging literature on comparative SLB performance has also identified ‘structural loopholes’ in regard to the use of late-target dates and early call options by issuers (Haq & Doumbia, 2022; Kölbel & Lambillon, 2022).

An asset manager in a Norwegian asset management firm (I4) explains that the choice to invest in SLBs versus use-of-proceeds bonds also ‘depends on what investors are looking for’:

Because pure green investors won't look at SLBs … . A really high impact investor probably wouldn’t look at SLBs because … it’s a general corporate bond where you don’t actually know specifically what the funds have been used for so it’s very difficult to calculate your impact. (Leading analyst, EF, I3)

In this sense labels and standards work in both directions: acting as both a response to investors’ expressed needs, but also performing a future promise that can recruit investors. This analyst identifies transition bonds as a better means to access finance for non-green companies than SLBs, because of their clearer narrative around use of funds.

By early 2022, some commentators were claiming that transition bonds were ‘dead’, because of the ‘lure of sustainability-linked bonds’ combined with the lack of clarity over what could be qualified, and thus financed, under the transition label. This latter problem was often attributed to the ICMA’s failure ‘to provide the same level of detail’ as that found in the Green Bond Principles (Webb, 2022). Thus in 2021, only a ‘handful of issuers’ had ‘struggled to reach $4 billion’ in transition bonds, despite a healthy total sustainable debt issuance of more than $1 trillion (Webb, 2022). But these figures may be deceptive: the morality of the first batch of transition bonds had been questioned, so the use of the transition label as a device to provide qualification of certain activities had been compromised, but SLBs (with climate targets) and green bonds (climate mitigation) were still financing transition activities.

Further improvements to transition taxonomies as a calculative device have since improved the label’s ability to qualify assets for financing, as questions over the morality of the earlier issues become historical. In particular, the Platform on Sustainable Finance (PoSF) advocated an extension of the EU’s sustainability taxonomy to cover transitional or ‘intermediate’ activities in March 2022 (EF, 2022b). This classification uses traffic lights to ‘incentivize continuous improvements’, even retiring the concept of stranded assets, and ‘brown’ assets beyond the mark by introducing a category of transition finance for ‘always significantly harmful’ (ASH) activities to fund decommissioning (PoSF, 2022, pp. 7–8, 13, 25). Katie House, a senior asset manager with Affirmative Investment Management summarizes:

The concept of transition bonds has been around for a little while, but it seems to have not really taken off; and I think that’s because defining eligible use of proceeds for a transition product is quite complicated. There’s no clear guidance in the market for that … . From our point of view, whether we consider a security to be a credible labelled transition bond, relies on whether the stated use of proceeds are delivering transition and it’s not just a way of packaging incremental change and basically like greenwashing. (I2)

In her view, transition bonds have struggled to define, measure and verify their environmental contribution, which has dampened their growth.

Measuring transition and impact

The PoSF taxonomy defines transition as a dynamic, change-demanding category, which differentiates the device, giving it utility and worth, in relation to other fixed-income labels which qualify activities using a steady state association between an asset and its ‘green’ characteristics (e.g. green bonds). However, this is not an absolute distinction, as the ‘mitigation’ category of green bonds also invites transition, while SLB KPIs can target change and/or static assets. The relative worth of different investments often involves a difficult comparison between a ‘steady state’ asset or a ‘transition/change’, not captured in simply a comparison of products per se. For example, investing in a pure play green asset, such as a solar renewables project, represents an additionality in the energy landscape, whereas investing in a sulphur extraction chimney at a coal-fired power station represents an emissions reduction in a high-emitting asset, but with a potentially larger contribution to overall emissions reduction. However, the morality of the latter investment will widely be seen as less. In ‘emerging markets’ investors often note the ‘transformational’ impact of their finance in companies ‘only at the beginning of their sustainability journey’, as compared to investing with ‘seasoned issuers’ in ‘established markets’ (S10).

Notably not present in the industry discussions we reviewed was any mention of litigation risk for mis-labelling of ‘green’, or of litigation in respect of tail risks if transition were not pursued. Participants also rarely referred to measurement of carbon emissions reductions in discussions of the worth of product labels. However, there was some fear of public ‘exposure’ if the instruments declared green were not subsequently perceived as green: investors were afraid ‘to be on the front page of the FT’ (I9).

Investment decision-making

There is a balance in investment decision-making between financial returns and green value, or the ‘greenium’ that rewards issuers with cheaper credit, but is income forgone by investors in return for their moral commitment (see Perkins, 2021). The greenium is the difference in spread between a green bond and a conventional bond of the same issue and is said to be significant and positive in public green bonds from 2014 to 2021 (Hinsche, 2021). Over the period of this research, green bonds were generally outperforming vanilla bonds, as they had less exposure to fossil fuel assets, aviation and transport, whose profits were reduced by COVID-19, such that investors gave up no yield potential (S10). However, the interest rate set on a green bond reflects the generic creditworthiness of the issuer, rather than any direct relationship to the profitability or otherwise of the underlying green assets. For example, the UK Treasury Green Gilt was given an ‘AA Outlook Stable’ by Standard & Poor’s (HM Treasury and UK Debt Management Office, 2021), which is basically a judgement of the UK government’s ability to repay. This means that questions should continue over the moral issue of whether the issuer should enjoy cheaper credit, and the investor commensurate lower returns than from a vanilla bond from the same issuer, when the public good aspects of green financing are so well established (Senior fund manager, S15).

Our interviewees and webinar contributors had different views on the relative importance of financial and environmental returns, summarized as (1) money-making takes priority; (2) a mix of both monetary incentives and green concerns; and (3) sustainability takes priority. For example, several interviewees saw investor returns as sacrosanct and paramount, where ‘we owe our investors a fiduciary duty to achieve the promised yield’ (I9). One interviewee, an investor partner in a renewable energy business, eschewed the green label despite his company’s pure play profile. His case was that renewables were more profitable than fossil fuels (always, if government subsidies for the latter could be removed), and that ‘green bond’ issuers were ‘not serious’, and ‘playing’ while other companies were getting [profitable] ‘infrastructure in the ground’ (I18). The Stock Exchange Board member also noted that while investor demand for green products was strong, the extent to which they were prepared to forgo return was unclear (I16). Fiduciary duty was still paramount, and ‘you start to stray into fairly problematic territory when you say your purpose must be all sorts of other things as well’ (I16). This view is also supported by recent academic research (e.g. Christophers, 2019).

But most descriptions of the investment process refer to the second view on the relationship between monetary incentives and green concerns; that these can exist in a notional balance or mixture. For example, a senior asset manager summarized that many green companies do not issue green bonds, while other companies see them as ‘very good add-on signals’ to obtain easier access to capital. The job of asset management is to assess the company’s overall strategy, and ‘look at product quality: we have a mandate to generate return. Credit score, spread and yield are key’ (S6, 2021).

Katie House (asset manager) speaks of balancing sustainability and profitability using separate credit and sustainability teams, where both review potential investments separately, with both having an effective veto, before it can ‘enter into our investable universe’ (I2).

Other fund managers claim that sustainability can take priority (position 3), since the ‘needs of the end-investors vary widely’ which generates different investor mandates within the same asset management firm. Some

have a looser mandate. For example, they merely have to exclude oil and gas firms in a process called negative screening … Other funds are targeted at more environmentally-conscious investors. These funds have a narrower mandate. For example, they are only able to invest money ring-fenced for direct impact on the transition to net zero. (I14)

If this is so, a range of differentiated labels would usefully respond to end-investors’ specified needs, roughly mapped to the ‘more or less green’ qualification of morality.

Discussion: Green labelling and market iterations

Some authors have argued that transition bonds developed in response to green bond markets locking out oil and gas issuers, despite some companies offering projects with green credentials (CBI, 2020, p. 3; Franklin et al., 2019). For example, Mafrig issued a transition bond after first failing to raise investor interest in a green bond. However, our research suggests that this barrier may not be surmountable by rebranding alone, since transition bonds have attracted similar criticism to green bonds – for example CBI’s critique of the subsequent Mafrig transition bond.

A related explanation, supported by most experts in our research, argued that transition bonds developed in order to reduce reputational risk in overall sustainability markets, by protecting the ‘purity’ of green bonds (BNP Paribas, 2019, p. 2) and reducing incentives for greenwashing (Deschryver & De Mariz, 2020, p. 17). However, a minority view was that ‘use of proceeds’ meant just that: finance should be extended to ‘brown’ industries, even under the green bond label, with risk managed by the firm rather than by market labelling. But complexities remain as risk evaluation and product labelling inevitably entrain entities/activities and industry/firm/sector risk simultaneously.

Transition bond advocates also argue that they facilitate a clear ‘narrative’ of transition, with ‘brown issuers … compelled to tell an articulated story’ (CBI, 2020, p. 15), to ‘clearly communicate what climate transition means in the context of their current business model and their future strategic direction’ (AXA Investment Managers, 2019). Similarly, Shrey Kohli, head of debt capital markets and funds at the London Stock Exchange Group, argued that investors in transition bonds care about how companies ‘articulate strategies that lead to net zero’ asking ‘where are we going to be in 2025 or 2035?’ (Wright, 2021). A scientifically verified, future-oriented ‘story’ or ‘journey’ of transition increasingly frames the comparative morality of transition bonds, while contributing the keystone of valuation, classification and certification processes. Label qualification consisted of a combination of calculations of emissions reduction across time -using more or less scientific proof – combined with a varying degree of moral signalling.

In general, green bond narratives recruit a positive, affective response in the ‘living now’, while transition bond narratives have a greater commensurability with ideas of changing and future-building. But for both products, issuers and investors expressed strong emotive allegiance to ‘action’ and ‘solutions’ in the context of ‘time running out’, sometimes self-identifying as society’s most logical and equipped responders to climate change.

The contested politics and science of classification

Alongside narratives, all green-inflected bonds are also valued using mainstream investment and valuation techniques, such as discounted cashflow (see Christophers, 2019; Langley et al., 2021). They are rated according to the full balance sheet and earnings potential of the issuer, rather than directly to the profitability of the ‘use-of-proceeds’ activity. This generates a theoretical conundrum. If all classifications of green are additional (and subordinate) to financial calculation, why does it matter what a product is called? Langley et al. (2021, p. 505), found tensions between a ‘mainstream market modality of qualification’ for green bonds (portfolio devices to measure expected returns) and an ‘ethical modality of qualification’ [where green additionality is subject to further, non-financial, often carbon-based measurement]. These tensions mean that capital is decarbonizing in ‘uncertain and incomplete ways’, making ‘limited progress’ (Langley et al., 2021, pp. 494, 511). We concur, but also point to the performativity of the current politics of classification: each competing concept borrowed from environmental discourse has prescient and differing consequences for future markets and societal distribution of risk and well-being. In short, for climate justice.

In this regard, we found several ambiguities that deserve future research attention in the way green worth and transition narratives are constructed as credible, particularly in respect to the referencing of regulatory regimes and ‘science’. Public and private regulatory regimes produce standards, taxonomies and climate risk disclosure instruments which act as calculative devices, in turn producing varying degrees of symbolic care or ‘feel good’ value, and sometimes material change. Green investment narratives then reference these regulatory regimes, with a selective adoption and folding in of the ‘science’ within them – for example in carbon accounting, paradigms of ‘Net Zero’, ‘Paris 1.5% alignment’, or more rarely ‘nature-based solutions’ or ‘nature positive’. As issuers and investors use the authority of science in product narratives and visions of possible futures, they borrow vocabulary (concepts, terms, indicators) from environmental scientists, which inevitably entangles the metrology and ontology of finance with that of natural science in a creative but difficult tension or mutual re-worlding (see Sullivan, 2018). Again, in so far as the future of green markets and green governance are discursively aligned, sharing good science, new opportunities could arise for effective, coordinated climate action.

However, the role and use of scientific knowledge and scientific experts in climate change governance is selective and problematic (c.f. Brown, 2009; Hulme, 2010; Turnhout et al., 2016), with tensions reproduced in green finance. In our transcripts, the apparent reliance on the authority of science is pervasive, but the material verification of environmental objectives is scant – for example of actual GHG reduction. This proof threshold differs from the veracity required of generic scientific method. Of course, issuers provide documents outlining the expected impact of the bonds, and investors claim to look at them carefully, with increasing focus on monitoring and reporting, but there is a difference between inputting numbers in a spreadsheet – i.e. the act of reporting – and demonstrating what these numbers mean with science. For example, in a study on marine biodiversity offsetting in Australia, Niner and Randalls (2021) suggest that when establishing offsets measures what matters is not the numbers themselves but rather the existence of a system of audit that gives, and is used, to gain authority. Similarly in our case of environmentally themed bonds, issuers and investors agree on processes of monitoring and reporting, with scientific vocabulary used to give legitimacy and authority to green narratives, yet there is less evidence of the types of place-based, temporally extensive scientific methods that would prove outcomes. The ontological translation between the world views of green financiers and environmental scientists ends in compromise: while the scientific method within climate change science requires robust and sometimes complex systems of proof, issuers and investors deem such high accuracy as too costly, unfeasible or unnecessary.

A ‘good enough’ version prevails (see Niner & Randalls, 2021), where scientifically referenced standards bring ‘just about enough’ assurance, to make bonds ‘legitimately tradable’ (Stephan, 2012, p. 626) in a ‘lenient zone of qualification’ (Perkins, 2021, p. 2044) Some asset managers spoke of the difficulties of accurately counting carbon emissions reductions in complex ecological systems; of the attribution problem of how investments affect these; and of how, even if accurate, these calculations should affect the price and reward of financial products. But also ‘science’ vocabulary sometimes appeared simply in a dramaturgical sense to signal and legitimate a virtuous financial world performed frontstage – without a backstage of actual computation. Here ‘science’ is ritualized, contained and institutionalized (c.f. Turnhout et al., 2016).

Conclusion

We have examined how concepts from environmental science – transition, green, sustainable – have shaped product experimentation in green finance, acquiring shifting socio-technical meanings within the politics of classification. Transition bond advocates suggest that they can protect the credibility, reputation, and therefore value, of green bonds, while simultaneously providing finance to ‘brown’ industries who are nonetheless intent on change to lower emitting outcomes. Advocates also value transition narratives, which increasingly reference ‘just [+ ‘transition’]’, ‘Net Zero’ or ‘Paris alignment’, promising planned and progressive carbon emissions reductions, in contrast to a steady state green investment. Detractors of transition finance point to their incrementalism and potential lengthening of the lifespan of dirty assets, in comparison to green bonds’ ability to build a new world ‘in the now’, or the flexibility of SLBs. But our research found incomplete and unstable classification systems and continuing contested modalities which reflect wider societal disagreements about who manages, how, and in what time frame, the pathway to the future.

Green finance markets are still forming, so the eventual product labels that become industry norms may yet change. However, there is evidence of several competing marketization strategies in green market-making: ‘pure’ qualification, where transition bonds associated with dirty assets might falter; distancing, where transition bonds usefully remove dirty assets from green bond investments; and dramaturgical, where scientific verification is absent, and narratives favour abstract credibilization and performance.

Ethical approval statement

We confirm that prior to starting the study, ethical approval was obtained from King’s College London reference number MRA-20/21-23353.

Acknowledgements

The authors would like to thank the editors and anonymous reviewers for their extremely helpful and insightful comments and suggestions.

Disclosure statement

The authors report there are no competing interests to declare.

Additional information

Funding

We gratefully received financial support from the King’s College London, Faculty of Social Science and Public Policy Research Fund (2020–2021)

Notes on contributors

Sarah Bracking

Sarah Bracking is a Professor in Climate and Society at King’s College London with research interests in climate and green finance. She is editor of Corruption and Development (2007), author of Money and power (2009), The financialisation of power (2016), and joint author of Valuing development, environment and conservation: Creating values that matter (2019).

Maud Borie

Maud Borie is a Lecturer in Environment, Science & Society at King’s College London with expertise in Science & Technology Studies. Her research focuses on resilience and disaster risk, and the politics of environmental knowledge, particularly in the context of biodiversity-related issues and in the green finance industry.

Glenn Sim

Glenn Sim is a PhD student at King's College London, exploring how banks and fund managers are incorporating climate risk into asset pricing. He holds a MA in Social Policy from the LSE and a BA(Hons)/MA(Oxon) in Geography from Oxford.

Theo Temple

Theo Temple is a PhD student at King's College London, examining the political geographies of property taxation in Brazil. He is interested in how critical work on public finance can contribute to debates about sovereign power and biopolitics. He holds a BA and MA from the University of Liverpool.

Notes

1 The term ‘brown bonds’ is widely used in green finance, where ‘brown’ refers to high-carbon emitting assets, industries or sectors. ‘Vanilla bonds’ are mainstream products with no carbon assignation, while ‘green’ are low-carbon emitting. Because of racialized connotation we have used ‘high-emitting’ to refer to these assets, industries or sectors, excepting in the text reporting on the research transcripts, where we have reproduced the language use of respondents.